IC Value
About Us
Editorial Board
Contact Us

Estimating The Survival Time Of HIV/AIDS Patients By Reducing The Number of Predictors Using Canonical Correlation Technique

Authors:Gurprit Grover, Anurag Sharma
Int J Biol Med Res. 2019; 10(2): 6714-6719  |  PDF File


This research is aimed atestimating the survival time of HIV/ AIDS patients by reducing the number of predictors using canonical correlation technique. In the first step, the survival times are estimated using AFTM taking all the available predictors. Then the set of available predictors is divided into two groups, viz., one group with patients' life style factors such as Age, Sex, Drugs, Alcohol, Smoking and Opportunistic Infection and the other group with socio- demographic factors such as State, living status of patients, their occupations, marital status and condition of their spouse's health. Then to determine the association between these two groups, canonical correlation technique is applied. If there is association between two groups of predictors, then by means of loadings and cross- loadings, the most significant predictors among the two groups are chosen. Lastly. AFT model is applied again to estimate the survival time in the presence of only selected significant predictors. It is observed that the survival times are not affected by this reduction of predictors and the estimates so obtained are more efficient as they have reduced standard error.